Tag
This paper presents a Mass-Conserving Perceptron (MCP) framework that reformulates conceptual hydrologic models into physically constrained, interpretable neural networks, evaluated across 513 CAMELS-US basins for snow-water representation.